gpt2

This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.6530

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.93	7	8.2816
9.7888	2.0	15	6.6539
8.253	2.93	22	5.9839
7.4098	4.0	30	5.5296
7.4098	4.93	37	5.1792
6.6836	6.0	45	4.8581
6.2698	6.93	52	4.6282
5.8092	8.0	60	4.4243
5.8092	8.93	67	4.2668
5.3803	10.0	75	4.1214
5.3501	10.93	82	4.0024
5.1278	12.0	90	3.8835
5.1278	12.93	97	3.8106
4.9471	14.0	105	3.7422
4.9279	14.93	112	3.7098
4.8129	16.0	120	3.6740
4.8129	16.93	127	3.6601
4.7258	18.0	135	3.6535
4.8643	18.67	140	3.6530

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2